1. Lecture 17 Ethernet r Widely deployed because: m First LAN technology m Simpler and less expensive than token LANs and ATM m Kept up with the speed race: 10, 100, 1000 Mbps m Many E-net technologies (cable, fiber etc). But they all share common characteristics

2. Lecture 17-2 Ethernet Frame Structure r Sending adapter encapsulates an IP datagram in Ethernet Frame which contains a Preamble, a Header, Data, and CRC fields r Preamble: 7 bytes with the pattern 10101010 followed by one byte with the pattern 10101011; used for synchronizing receiver to sender clock (clocks are never exact, some drift is highly likely) r Header contains Destination and Source Addresses and a Type field r Addresses: 6 bytes, frame is received by all adapters on a LAN and dropped if address does not match r Type: indicates the higher layer protocol, mostly IP but others may be supported such as Novell IPX and AppleTalk) r CRC: checked at receiver, if error is detected, the frame is simply dropped

3. Lecture 17-3 CSMA/CD A: sense channel, if idle then {transmit and monitor the channel; If detect another transmission then { abort and send jam signal; update # collisions; delay as required by exponential backoff algorithm; go to A } else {done with the frame} } else {wait until ongoing transmission is over and go to A}

4. Lecture 17-4 CSMA/CD (Cont.) r Jam Signal: to make sure all other transmitters are aware of the collision; 48 bits; r Exponential Backoff: m Goal is to adapt the offered rate by transmitters to the estimated current load (ie backoff when load is heavy) m After the first collision Choose K from {0,1}; delay is K x 512 bit transmission times m After second collision choose K from {0,1,2,3}… m After ten or more collisions, choose K from {0,1,2,3,4,…,1023} r Under this scheme a new frame has a chance of sneaking in the first attempt, even in heavy traffic r Ethernet Efficiency: under heavy traffic and large number of nodes:

5. Lecture 17-5 Ethernet Technologies: 10Base2 r 10=>10Mbps; 2=>under 200 meters maximum length of a cable segment; also referred to as “Cheapnet” r Uses thin coaxial cable in a bus topology r Repeaters are used to connect multiple segments (up to 5); a repeater repeats the bits it hears on one interface to its other interfaces, ie a physical layer device only!

6. Lecture 17-6 10BaseT and 100BaseT r 10/100 Mbps rate; latter called “fast ethernet” r T stands for Twisted Pair r Hub to which nodes are connected by twisted pair, thus “star topology” r CSMA/CD implemented at the Hub r Max distance from node to Hub is 100 meters r Hub can disconnect a “jabbering adapter”; 10base2 would not work if an adapter does not stop transmitting on the cable r Hub can gather monitoring information and statistics for display to LAN administrators

7. Lecture 17-7 Gbit Ethernet r Use standard Ethernet frame format r Allows for Point-to-point links and shared broadcast channels r In shared mode, CSMA/CD is used r Full-Duplex at 1 Gbps for point-to-point links

8. Lecture 17-8 Hubs, Bridges and Switches r Used for extending LANs in terms of geographical coverage, number of nodes, administration capabilities, etc. r Differ in regards to: m collision domain isolation m layer at which they operate

9. Lecture 17-9 Hubs r Physical Layer devices: essentially repeaters operating at bit levels: repeat received bits on one interface to all other interfaces r Hubs can be arranged in a hierarchy (or multi-tier design), with a backbone hub at its top r Each connected LAN is referred to as a LAN segment r Hubs do not isolate collision domains: a node may collide with any node residing at any segment in the LAN

10. Lecture 17-10 Hubs (Cont.) r Hub Advantages: + Simple, inexpensive device + Multi-tier provides graceful degradation: portions of the LAN continue to operate if one of the hubs malfunction + Extends maximum distance between node pairs (100m per Hub) + Interdepartmental Communication r Hub Limitations: - Single collision domain results in no increase in max throughput; the multi-tier throughput same as the the single segment throughput - Cannot connect different Ethernet types (eg 10BaseT and 100baseT)

11. Lecture 17-11 Bridges r Link Layer devices: operate on Ethernet frames, examining the frame header and selectively forwarding a frame based on its destination r Bridge isolates collision domains since it buffers frames r When a frame is to be forwarded on a segment, the bridge uses CSMA/CD to access the segment and transmit r Bridge advantages: + Isolates collision domains resulting in higher total max throughput, and does not limit the number of nodes nor geographical coverage + Can connect different type Ethernet since it is a store and forward device + Transparent: no need for any change to hosts LAN adapters

12. Lecture 17-12 Backbone Bridge

13. Lecture 17-13 Interconnection Without Backbone r Not recommended for two reasons: - Single point of failure at Computer Science hub - All traffic between EE and SE must path over CS segment

14. Lecture 17-14 Bridge Filtering r Bridges learn which hosts can be reached through which interfaces and maintain filtering tables (bridge tables) r A filtering table entry: (Node LAN Address, Bridge Interface, Time Stamp) where Node LAN Address is the 6 byte physical address r Filtering procedure: if destination is on LAN on which frame was received then drop the frame else { lookup filtering table if entry found for destination then forward the frame on interface indicated; else flood; /* forward on all but the interface on which the frame arrived*/ }

15. Lecture 17-15 Bridge Learning r When a frame is received, the bridge “learns” from the source address and updates its filtering table (Node LAN Address, Bridge Interface, Time Stamp) r Stale entries in the Filtering Table are dropped (TTL can be 60 minutes)

16. Lecture 17-16 Bridges Spanning Tree r For increased reliability, it is desirable to have redundant, alternate paths from a source to a destination r With multiple simultaneous paths however, cycles result on which bridges may multiply and forward a frame forever r Solution is organizing the set of bridges in a spanning tree by disabling a subset of the interfaces in the bridges: Disabled

17. Lecture 17-17 Bridges Vs. Routers r Both are store-and-forward devices, but Routers are Network Layer devices (examine network layer headers) and Bridges are Link Layer devices r Routers maintain routing tables and implement routing algorithms, bridges maintain filtering tables and implement filtering, learning and spanning tree algorithms

18. Lecture 17-18 Routers Vs. Bridges (Cont) r Bridges + and - + Bridge operation is simpler requiring less processing bandwidth - Topologies are restricted with bridges: a spanning tree must be built to avoid cycles - Bridges do not offer protection from broadcast storms (endless broadcasting by a host will be forwarded by a bridge) r Routers + and - + Arbitrary topologies can be supported, cycling is limited by TTL counters + Provide firewall protection against broadcast storms - Require IP address configuration (not plug and play) - Require higher processing bandwidth r Bridges do well in small (few hundred hosts) while routers are required in large networks (thousands of hosts)

19. Lecture 17-19 Ethernet Switches r A switch is a device that incorporates bridge functions as well as point-to-point ‘dedicated connections’ r A host attached to a switch via a dedicated point-to-point connection; will always sense the medium as idle; no collisions ever! r Ethernet Switches provide a combinations of shared/dedicated, 10/100/1000 Mbps connections r Some E-net switches support cut-through switching: frame forwarded immediately to destination without awaiting for assembly of the entire frame in the switch buffer; slight reduction in latency r Ethernet switches vary in size, with the largest ones incorporating a high bandwidth interconnection network

20. Lecture 17-20 Ethernet Switches (Cont) Dedicated Shared